lofield.fm
lofield.fm lets you create lofi beats by describing what you want in plain English. No music theory or DAW experience needed — just type a description and the AI turns it into playable music running in your browser.
How it works
You describe the sound you're after in natural language. The AI converts your prompt into executable Tone.js code, which runs in the browser to produce real-time audio. You can refine beats iteratively through conversation, adjusting the generated code or tweaking performance controls until it sounds right.
Key features
- Text to music — describe a vibe and get playable Tone.js code
- Performance controls — tempo (60–200 BPM), swing, filters, reverb, and delay, all adjustable in real time
- Multi-layer composition — independent drum, bass, melody, and pad tracks
- Automation recording — capture parameter changes during playback
- Version history — every revision tracked with diffs and one-click restore
- Export options — WAV rendering, shareable public links, and raw code export
- Code editor — CodeMirror 6 with syntax highlighting and error detection
- Visual timeline — 32-bar view with playback position and section markers
Stack
Next.js, React 19, and Tailwind CSS on the frontend. Tone.js for audio synthesis. OpenAI GPT-4o via Vercel AI SDK for prompt-to-code generation. Supabase for auth, database, and storage. CodeMirror 6 for the code editor.